Abstract Background and Objectives Serratia marcescens is an emerging nosocomial pathogen, and the carbapenemase blaNDM has been reported in several surveys in Romania. We aimed to investigate the molecular epidemiology of S. marcescens in two Romanian hospitals over 2010–15, including a neonatal NDM-1 S. marcescens outbreak. Methods Isolates were sequenced using Illumina technology together with carbapenem-non-susceptible NDM-1-positive and NDM-1-negative Klebsiella pneumoniae and Enterobacter cloacae to provide genomic context. A subset was sequenced with MinION to fully resolve NDM-1 plasmid structures. Resistance genes, plasmid replicons and ISs were identified in silico for all isolates; an annotated phylogeny was reconstructed for S. marcescens. Fully resolved study NDM-1 plasmid sequences were compared with the most closely related publicly available NDM-1 plasmid reference. Results 44/45 isolates were successfully sequenced (S. marcescens, n = 33; K. pneumoniae, n = 7; E. cloacae, n = 4); 10 with MinION. The S. marcescens phylogeny demonstrated several discrete clusters of NDM-1-positive and -negative isolates. All NDM-1-positive isolates across species harboured a pKOX_NDM1-like plasmid; more detailed comparisons of the plasmid structures demonstrated a number of differences, but highlighted the largely conserved plasmid backbones across species and hospital sites. Conclusions The molecular epidemiology is most consistent with the importation of a pKOX_NDM1-like plasmid into Romania and its dissemination amongst K. pneumoniae/E. cloacae and subsequently S. marcescens across hospitals. The data suggested multiple acquisitions of this plasmid by S. marcescens in the two hospitals studied; transmission events within centres, including a large outbreak on the Targu Mures neonatal unit; and sharing of the pKOX_NDM1-like plasmid between species within outbreaks. Introduction Carbapenemase-producing Enterobacteriaceae (CPE) have emerged post-1990 and spread worldwide, representing a serious clinical threat.1 Although carbapenem resistance can arise through various mechanisms, acquisition of genes (blaKPC, blaNDM-1, blaOXA-48) encoding Klebsiella pneumoniae carbapenemase (KPC), New Delhi metallo-β-lactamase (NDM) and OXA-48 carbapenemase, respectively, are currently of most concern. These genes are readily transmitted intra- and inter-species via mobile genetic elements that have facilitated their worldwide dissemination.2 Compared with Escherichia coli and K. pneumoniae, two of the most frequently identified CPE species, Serratia marcescens has been less commonly associated with any of the major transmissible carbapenemases, although acquisition of blaKPC has been described recently in China, Greece and the United States.3–5,S. marcescens has also traditionally been regarded as an opportunistic pathogen causing disease in specifically vulnerable populations, such as neonates.6,7 However, Serratia spp. have become increasingly adapted to hospital environments and have emerged as important agents of hospital-acquired infections affecting all age groups.8–10 The inherent resistance of Serratia spp. to multiple antimicrobial classes may have facilitated this adaptation, and leaves few therapeutic options open for treating carbapenem-resistant S. marcescens infections.11–13 blaNDM-1 has been increasingly reported in patients with healthcare exposure on the Indian subcontinent or in the Balkan region.1,14,15 The occurrence of blaNDM-1 in S. marcescens appears rare, with only a handful of isolates identified in Germany,16 the UK17 and Egypt.18 However, a recent faecal screening/surveillance project for carbapenemase-producing Gram-negative bacteria in three Romanian hospitals identified 17 NDM-1 S. marcescens in clinical, environmental and screening isolates,19 all of which were from the same neonatal unit at a single hospital, consistent with an outbreak. PCR-based plasmid analyses identified blaNDM-1 in association with conjugative IncFII plasmids in both S. marcescens and K. pneumoniae, suggesting inter-species transfer of blaNDM-1, possibly from K. pneumoniae into a well-established hospital-adapted S. marcescens clone. In this study, we used Illumina (short-read) and MinION (long-read) WGS of NDM-1- and non-NDM-1 S. marcescens and other Enterobacteriaceae isolates obtained from: (i) the faecal screening/surveillance study; (ii) a previous outbreak and sporadic S. marcescens infections in this neonatal unit; and (iii) a second hospital in an adjacent county, to investigate the evidence for nosocomial and inter-hospital dissemination of NDM-1 S. marcescens and the transmission of NDM-1 plasmids amongst Enterobacteriaceae in this context in Romania. Materials and methods Hospital settings Bacterial isolates were collected from two teaching hospitals located in central/north Romania: one in Targu Mures (1099 beds, ∼42 000 patients/year), and one in Cluj-Napoca (252 beds, ∼20 000 patients/year). In Targu Mures, the neonatology unit had 65 beds [10 were for neonatal intensive care (NICU), 15 for the care of premature neonates]. Overview of S. marcescens outbreaks in Targu Mures hospital During April–June 2010, 11 neonates were found to be colonized (faecal/pharyngeal) with carbapenem-susceptible S. marcescens, 4 of whom (36%) developed S. marcescens bacteraemia. Environmental screening did not identify any source; PFGE (data not shown) was consistent with transmission of a clonal strain with the main vector presumed to be colonized neonates. The main contributing factors to carbapenem-susceptible S. marcescens dissemination were thought to be overcrowding in the NICU and selective pressure exerted by ampicillin/sulbactam and gentamicin prophylaxis given to high-risk neonates (premature and/or complicated delivery). All patients survived and the outbreak was declared over in July 2010. Sporadic infections with carbapenem-susceptible S. marcescens recurred in 2012, whilst both colonization events and infections with carbapenem non-susceptible-K. pneumoniae and carbapenem non-susceptible E. cloacae occurred during 2012–13 in the NICU and paediatric cardiology ward. Following this, in August 2014, several cases of carbapenem non-susceptible S. marcescens pharyngeal colonization occurred in the neonatology unit and an outbreak was declared in October 2014; enhanced infection control measures were implemented, including faecal screening of all admissions/transfers to the unit. At this point, the genetic mechanism mediating carbapenem resistance was shown by PCR to be blaNDM-1. Six cases of NDM-1 S. marcescens bacteraemia occurred October–December 2014; ongoing NDM-1 S. marcescens faecal/pharyngeal colonization cases were identified up until May 2015. In both outbreaks in the Targu Mures NICU (2010, 2014–15), infection control measures were reinforced, consisting initially of isolating colonized/infected patients and their contacts, terminal cleaning with quaternary ammonium compounds, and special attention to NICU disinfection. At the same time, hospital staff were re-educated with respect to contact precautions and hand hygiene procedures to try and reduce the risk of any staff–patient transmission. Isolates Targu Mures hospital isolates Isolates from Targu Mures included those obtained from faecal screens performed during active carbapenemase surveillance in the 2014–15 S. marcescens outbreak, as well as any clinical carbapenem-non-susceptible S. marcescens isolated from the neonatal unit. Other carbapenem-non-susceptible S. marcescens, carbapenem-non-susceptible K. pneumoniae and carbapenem-non-susceptible E. cloacae obtained from clinical specimens on other hospital units were also included for comparison. Carbapenem-non-susceptible S. marcescens cultured from environmental screening samples (October 2014–April 2015); previous NDM-1 K. pneumoniae and NDM-1 E. cloacae isolated from clinical specimens; and carbapenem-susceptible S. marcescens obtained from a previous outbreak in the same neonatal unit (2010) and from sporadic clinical cases (2012) were also included in the study. Ethics approval for surveillance was obtained from the Hospital Research Ethics Committee (#138; 23/12/2013). Cluj-Napoca hospital isolates Cluj-Napoca is a city neighbouring Targu Mures (∼80 km away from Targu Mures) and patients travel between these cities for medical intervention not provided locally. A subset of stored S. marcescens isolates with a carbapenemase-producing phenotype (ertapenem- and meropenem-resistant and CarbaNP test-positive) from urine samples were randomly selected to identify evidence for inter-hospital transmission. Microbiological methods For faecal screening, a modified method from the US CDC involving enrichment and subculture on selective media [trypticase soy broth (TSB)/MacConkey + carbapenem] was used.20 In brief, a rectal swab or a small amount of faecal sample was placed in 5 mL of TSB containing a meropenem (10 μg) disc and incubated at 35 ± 2°C overnight. 100 μL of inoculated TSB was subsequently streaked onto MacConkey agar and carbapenem discs (ertapenem 10 μg and meropenem 10 μg) were placed on different sectors of the inoculum (all discs and media from Oxoid, Basingstoke, UK); screening cut-off values for CPE were applied according to the EUCAST methodology (v5.0, available at: http://www.eucast.org/fileadmin/src/media/PDFs/EUCAST_files/Breakpoint_tables/v_5.0_Breakpoint_Table_01.pdf). Presence of carbapenemases was confirmed phenotypically using the CARBA NP test (BioMérieux, Paris, France) at Cluj-Napoca or by PCR (for blaNDM, blaKPC, blaIMP and blaVIM) at Targu Mures.21,22 Clinical specimens were processed according to local protocols. Species identification was performed using the VITEK 2 (BioMérieux); susceptibility testing was performed by microbroth dilution (TREK Diagnostic Systems, West Sussex, UK) for isolates from the faecal surveillance study and by disc diffusion for the others (EUCAST guidelines). DNA extractions and Illumina/MinION sequencing DNA extraction of sub-cultured isolates for Illumina sequencing was performed using the Quickgene commercial kit (Fujifilm, Tokyo, Japan), with an additional mechanical lysis step following chemical lysis (FastPrep, MP Biomedicals, Santa Ana, USA). Isolates were sequenced using the Illumina HiSeq 2500 (150 bp paired-end reads; sequencing coverage ∼100×). Ten isolates were selected for MinION sequencing based on Illumina data; isolates were cultured from frozen stocks (−80 °C) on Columbia Blood agar in the presence of three ertapenem discs (10 μg; Oxoid, Thermo Fisher Scientific, USA) overnight at 37 °C. DNA was then extracted using the Qiagen Genomic tip 100/G kit as per the manufacturer’s instructions (Qiagen, Venlo, the Netherlands), quantified using the Qubit 2.0 Fluorometer (Life Technologies, USA), and fragment lengths assessed using the TapeStation 2200 (Agilent, UK). Library preparation was performed using the 2D native barcoding protocol, targeting 8 kb fragments, and with an input of 3–4 μg DNA. All steps were performed following the manufacturer’s protocol with native barcoding to multiplex two samples per sequencing run (EXP-NBD002 and SQK-LSK208). Libraries were sequenced using the best available flow cells at the time (R9.4). Libraries were topped up after 12 h, and sequencing performed for 48 h in total. Data were base-called in real time via Metrichor, using the best available workflow at the time (v1.125; November–December 2016 and January 2017; ONT, UK). Insufficient sequence data yields were generated for two samples (multiplexed in a single run; 7209 and 18ES), requiring repeat sequencing of the same DNA extract by the same methods. The sequence data from both runs were merged to generate a final set of data for further processing. WGS analysis For Illumina data, read quality- and adapter-related trimming and filtering were performed using BBDuk (BBTools; available at https://jgi.doe.gov/data-and-tools/bbtools/), and species identification was confirmed in silico on the basis of the top species match identified using Kraken.23 The phylogeny of S. marcescens isolates was reconstructed using IQTree24 from a consensus fasta of variant sites generated from mapping to the S. marcescens reference WW4 (NC_020211.1). Read mapping was performed using Stampy,25 with mapping and filtering of variant calls based on a set of thresholds related to base and mapping quality and read coverage, as previously described.26 For phylogenetic reconstruction, a General Time Reversible model27 was used with a gamma category allowing variable mutation rates across sites, and a maximum parsimony starting tree. The maximum-likelihood phylogeny output by IQTree was then corrected for recombination using ClonalFrameML28 (default parameters), and visualized in iTOL.29 De novo assemblies of short-read sequenced isolates were generated with SPAdes (v3.6; default settings).30 We used an arbitrary assembly size of >6.5 Mb as a threshold to identify any mixed sequences (including mixtures of the same species not identified by Kraken) and remove them from the analysis. In silico multi-locus sequence-typing (MLST) for K. pneumoniae and E. cloacae isolates was performed using BLASTn of the sequence assemblies against the MLST allele databases of these two species downloaded from https://pubmlst.org/general.shtml. Plasmid replicon typing was performed using BLASTn against the PlasmidFinder database,31 and IS typing using the ISFinder database.32 Resistance gene characterization was performed using the in-house script ResistType, a BLASTn/local reassembly-based approach for querying the presence/absence of known chromosomal and acquired resistance mechanisms using a database as described in Stoesser et al.33 (scripts and database available at: https://github.com/hangphan/resistType_docker). This database includes >2000 variants from >70 resistance gene families, covering all the major resistance gene mechanisms and many minor ones. NDM-1 plasmid typing was performed using a reference database of NDM-1 plasmids generated in this study and the subset harbouring NDM-1 from a curated database of plasmids downloaded from NCBI.34 BLASTn was then used to query the sequence assemblies for the presence of similar blaNDM-carrying plasmid structures. An isolate was characterized as having candidate NDM-1-plasmid ‘X’ if >90% of plasmid ‘X’ could be aligned with the query assembly’s contigs, where for each contig, >90% of the contig length matched plasmid ‘X’ with >95% sequence identity. Thresholds were set so that an isolate could have multiple matches to the NDM-1 plasmid database. Poretools35 was used to extract 2D reads from base-called MinION sequencing data and Canu (v1.4)36 to generate de novo assemblies, which were error-corrected using short-read Illumina data and the pilon assembly polisher (v1.18).37 A plasmid was defined as circularized/complete if it had >1 kb overlapping ends with >99% sequence identity. Alignments of fully reconstructed plasmid sequences were visualized and annotated in Geneious (version: R9).38 Data accession All short-read sequencing data (fastq files) and long-read sequencing data (assemblies) from the study are available in GenBank (BioProject accession: PRJNA396838). Results Bacterial isolates Four hundred and seventeen faecal/pharyngeal (n = 135), staff hand swabs (n = 90) and environmental samples (n = 192) were collected from the neonatology unit during the Targu Mures surveillance study, which overlapped with the major S. marcescens outbreak period (January–May 2015). From these, 24 carbapenem-non-susceptible Enterobacteriaceae were cultured (S. marcescens, n = 19, K. pneumoniae, n = 5; 19/19 S. marcescens were found by PCR to carry blaNDM-1; 5/5 K. pneumoniae were Carba NP test-positive but not tested by PCR). Positive cultures were obtained from faecal screens (n = 14) and pharyngeal secretions (n = 7) from 14 neonates/children, and three environmental samples (one from a bed sheet, two from feeding tubes). NDM-1 S. marcescens neonatal bacteraemia isolates from the 2014–15 outbreak (n = 5), carbapenem-susceptible S. marcescens from an earlier outbreak (2010, n = 3) and sporadic carbapenem-susceptible S. marcescens cases (2012, n = 3), as well as previous sporadic NDM-1 K. pneumoniae (n = 3, 2013) and NDM-1 E. cloacae (n = 3, 2013) were included. From Cluj-Napoca, five carbapenem non-susceptible S. marcescens were obtained for comparison. One sample was lost in processing, resulting in 45 sent for WGS (Table S1, available as Supplementary data at JAC Online). WGS data analysis Forty-four of the 45 sequences passed assembly quality control, including K. pneumoniae (n = 7; 2 new ST, 3 ST-15, 1 ST-45, 1 ST-307), E. cloacae (n = 4; 2 ST-90, 1 ST-254, 1 ST-104) and S. marcescens isolates (n = 33; 27 NDM-1-positive; Figure S1). The earliest confirmed NDM-1-positive strains of each species were S. marcescens 7209 (28/03/2013) in Cluj-Napoca and in Targu Mures E. cloacae 20ES (27/07/2013), K. pneumoniae 19ES (07/08/2013), and S. marcescens 12TM (22/10/2014). Phylogenetic analysis of S. marcescens isolates was consistent with a clonal outbreak of NDM-1 S. marcescens across clinical, environmental and faecal screening isolates in the Targu Mures neonatal/paediatric units [<25 single nucleotide variations (SNVs) distance across all isolates, 2014–15]. This clonal strain was however chromosomally distantly related (>3300 SNVs) to the 2010–12 NDM-1-negative S. marcescens isolates from Targu Mures and the 2013–15 NDM-1-positive S. marcescens isolates from Cluj-Napoca (Figure 1). Other antimicrobial resistance genes were common (Figure 1); two isolates, K. pneumoniae 17ES from Targu Mures and S. marcescens 9275 from Cluj-Napoca, harboured both blaNDM-1 and blaOXA-48. Figure 1. View largeDownload slide Phylogeny of the sequenced Serratia marcescens strains analysed in this study, annotated by location, sampling date and genomic features. Genetically related clusters of NDM-1-positive isolates are denoted by the letters A–C. Figure 1. View largeDownload slide Phylogeny of the sequenced Serratia marcescens strains analysed in this study, annotated by location, sampling date and genomic features. Genetically related clusters of NDM-1-positive isolates are denoted by the letters A–C. Long-read sequencing generated 12–98× coverage for the 10 isolates sequenced (Figure S2), confirmed the absence of blaNDM in one S. marcescens isolate (14ES, 2012), and enabled closure of the NDM-1 plasmid sequences (87.5–120 kb) in the other nine isolates [S. marcescens (n = 4), K. pneumoniae (n = 3) and E. cloacae (n = 2); Table S2]. All nine NDM-1 plasmids shared high sequence similarity with pKOX_NDM1 (accession number NC_021501.1), an IncFIIY plasmid first identified in Taiwan in 201039 (Figure 2). Several genetic changes with respect to pKOX_NDM1 were observed in the study plasmids, including: scattered nucleotide mutations; clusters of nucleotide mutations consistent with small recombination events; and small and large indels. Examples of indels included: (i) deletions of all or part of a cluster of genes including cph (encoding for a bacterial phytochrome), merR (regulator of the mer operon encoding for mercury resistance) and the capsular polysaccharide biosynthesis operon (including kpsC), the latter two of which have been associated with antimicrobial resistance, environmental persistence and/or virulence40,41 (pNDM_22ES, pNDM_7209, pNDM_9580); (ii) deletions of several open reading frames involved in conjugal transfer (pNDM_7209); and (iii) acquisition of a set of putative phage-associated elements in neonatal outbreak isolates from 2014 to 2015 (pNDM_12TM, pNDM_4TM, pNDM_5TM, pNDM_6TM) (Figure 2). A large recombination event was also observed in pNDM_5TM and pNDM_6TM, encompassing the region downstream of the putative phage to traM (Figure 2), and genetically highly similar (99%) to ∼18 kb of another reference plasmid, pCAV1344-78 (GenBank accession: CP011621, Charlottesville, USA). This region contained two klcA paralogues; klcA encodes for an anti-restriction protein, which has recently been shown to promote horizontal gene transfer amongst KPC plasmids.42 Figure 2. View largeDownload slide Alignment of study NDM-1 plasmid sequences and the reference NDM-1 plasmid pKOX_NDM1. Aligned bars adjacent to plasmid names represent plasmid sequences: light grey denotes regions with 100% sequence identity; black represents nucleotide diversity between sequences; and thin lines represent indels. Coding sequences are represented by fat arrows below individual sequence bars and are colour coded as per the colour key. The inset schematic describing genetic variation between sequences depicts examples of evolutionary events identified: (a) single nucleotide level change, (b) small indels (≤100 bp), (c) large indels (>100 bp), (d) recombination events. Sequence labels are coloured by location (grey, reference; black, Targu Mures hospital; blue, Cluj-Napoca hospital). The letters A–C correspond to isolate clusters in Figure 1, and represent the relatedness of the host strain in which the NDM-1 plasmids were found. Figure 2. View largeDownload slide Alignment of study NDM-1 plasmid sequences and the reference NDM-1 plasmid pKOX_NDM1. Aligned bars adjacent to plasmid names represent plasmid sequences: light grey denotes regions with 100% sequence identity; black represents nucleotide diversity between sequences; and thin lines represent indels. Coding sequences are represented by fat arrows below individual sequence bars and are colour coded as per the colour key. The inset schematic describing genetic variation between sequences depicts examples of evolutionary events identified: (a) single nucleotide level change, (b) small indels (≤100 bp), (c) large indels (>100 bp), (d) recombination events. Sequence labels are coloured by location (grey, reference; black, Targu Mures hospital; blue, Cluj-Napoca hospital). The letters A–C correspond to isolate clusters in Figure 1, and represent the relatedness of the host strain in which the NDM-1 plasmids were found. Plasmid typing of the other isolates (S. marcescens, K. pneumoniae and E. cloacae) using short-read Illumina data was consistent with pKOX_NDM1-like plasmids being present in all NDM-1-positive strains, although the context and structure of these plasmids could not be resolved from the Illumina assemblies. There were no significant matches to any non-pKOX_NDM1 NDM-1-harbouring plasmids (Figure S1; all NDM-1-harbouring isolates have a top match to pKOX_NDM1). Discussion Here, we show that the molecular epidemiology of carbapenemase-non-susceptible S. marcescens in two regional Romanian hospitals is consistent with the emergence of multiple NDM-1 S. marcescens strain clusters in conjunction with pKOX_NDM1-like plasmids, causing clinical outbreaks over 2013–15. NDM-1 S. marcescens was isolated from both asymptomatic neonates and a small number of environmental samples, and caused high rates of invasive disease (35%) in this susceptible population. Highly genetically similar pKOX_NDM1-like plasmids were observed in contemporaneously circulating E. cloacae and K. pneumoniae, suggestive of a multi-species, multi-hospital pKOX_NDM1-like plasmid transmission network, and acquisition of this plasmid by genetically distinct strains of S. marcescens in Targu Mures and Cluj-Napoca hospitals, causing the observed outbreak in Targu Mures. Staff and environmental sampling failed to identify a major non-patient reservoir, as elsewhere.10 Although most NDM-1 S. marcescens isolates were obtained in the context of the neonatal outbreak, several NDM-1 S. marcescens isolates were from clinical specimens obtained from adults, suggesting that these isolates are also important to consider in non-neonatal patient groups. There are limited data on carbapenemase-producing Enterobacteriaceae in Romania, but the earliest published reports of NDM-1-producing Enterobacteriaceae in this country are from Targu Mures hospital, one of our study hospitals. Surveillance between January 2010 and September 2012 identified 9/2317 (0.3%) of non-duplicate Enterobacteriaceae as carbapenemase producers, of which five harboured blaNDM-1 [E. cloacae n = 2, K. pneumoniae n = 2 (one with OXA-181), E. coli n = 1] and four blaOXA-48 (K. pneumoniae n = 3 and S. marcescens n = 1).43 Surveillance of 100 isolates from 2011 mediated as part of the global SENTRY antimicrobial resistance surveillance programme identified three CPE (3%), two of which were NDM-1 producers (both E. cloacae; one from Bucharest and one from Cluj-Napoca), and one an OXA-48 producer (K. pneumoniae; from Bucharest).44 A third study focusing on surveillance in Cluj-Napoca, August 2011–November 2013,45 similarly identified ∼3% CPE prevalence (64/1903 isolates tested; all in individuals >16 years of age), with the earliest NDM-1 isolate an E. cloacae (26/9/2011) obtained from a patient in ambulatory care, and the first NDM-1 S. marcescens isolate reportedly from February 2013. Although the sampling in each of these studies was geographically and temporally restricted, the data from these studies and our survey would suggest that blaNDM-1 was present in non-Serratia spp. first, and subsequently emerged in Serratia spp. Much of the recent surveillance of carbapenemase producers unsurprisingly focuses on those species of Enterobacteriaceae that cause most clinical disease, namely E. coli and/or K. pneumoniae. The most recent report of the European Antimicrobial Resistance Surveillance Network (EARS-Net) shows a significant increasing trend in carbapenem resistance in both E. coli and K. pneumoniae invasive isolates from Romania which is concerning.46 In addition, a recent European-wide survey (EuSCAPE) investigating clinical isolates of these two species suggested that in Romania most carbapenem resistance was attributable to clonal dissemination of a K. pneumoniae strain harbouring blaOXA-48.47 Our data and other recently published studies19,45,48,49 suggest that the epidemiology of carbapenemase producers is more complicated, and different mechanisms may be circulating in different species, with a degree of horizontal exchange of successful plasmids. In Czobor et al.,48 several NDM-1 isolates were harbouring IncFIIY plasmids, consistent with these being pKOX_NDM1-like, although with the limited plasmid typing performed this cannot be confirmed. Surveillance is therefore difficult, needs to encompass asymptomatic individuals and a wide range of species types over time in order to be accurate, and needs to be high resolution and comparable across studies. A limitation of our study is that our strain collection was not complete (i.e. did not include all NDM-positive isolates as well as consecutive non-NDM isolates representing background diversity for the two sites), but was predominantly focused on investigating NDM-1 strains collected as part of outbreak investigation, and sequencing of a small number of isolates that had been stored ad hoc. Although WGS is now relatively commonly used for molecular epidemiology and outbreak investigation, this is the first study to our knowledge to use WGS data to investigate the evidence for NDM-1 CPE transmission in Romania, and the mobile genetic elements involved in the acquisition and dissemination of blaNDM-1 in nosocomial S. marcescens. The use of long-read WGS data to fully resolve nine of the NDM-1 plasmid structures involved in this outbreak overcomes some of the pitfalls in investigating plasmid epidemiology using short-read WGS data, with which many plasmids are only reconstructed as partial fragments or contigs. However, evolutionary models to account for the genetic variation and plasticity observed in these plasmid structures (mutations, recombination, indels) have not yet been developed, and it was not possible to determine a clear sequence of plasmid transmission events by visual inspection alone, particularly in view of the fact that we are likely to have sampled only a proportion of potentially circulating plasmids. We also cannot exclude the possibility of multiple, discrete, pKOX_NDM1-like plasmid importation events into these hospitals. In this context, our phylogenetic analysis of bacterial strains and high-resolution analysis of nine NDM-1 plasmids in circulating K. pneumoniae, E. cloacae and S. marcescens would be consistent with the transfer of pKOX_NDM1 from major non-Serratia spp. into several S. marcescens host strains. We found no evidence of any non-pKOX_NDM1-like plasmid backgrounds in the 34 NDM-1-positive Enterobacteriaceae sequenced across the study timeframe and two hospitals, which would be most consistent with the importation of this plasmid type and its local dissemination, as part of a pKOX_NDM1-like plasmid-associated outbreak, as opposed to the emergence of a separate successful Romanian NDM-1 plasmid. Further studies exploring the global diversity of pKOX_NDM1-like plasmids, which have been reported in several Enterobacteriaceae isolates from Canada,50 China,51 Australia/New Zealand52 and Saudi Arabia53 would be useful to identify the extent of transmission of this global, multi-species NDM-1 vector. Acknowledgements The authors thank the clinical/laboratory staff and patients at both study hospitals. Funding This work was supported by the Health Innovation Challenge Fund [a parallel funding partnership between the Wellcome Trust (WT098615/Z/12/Z) and the Department of Health (grant HICF-T5-358)]; the National Institute for Health Research (NIHR) Oxford Biomedical Research Center (BRC) Program; and the Health Protection Research Unit (NIHR HPRU) in Healthcare Associated Infections and Antimicrobial Resistance at the University of Oxford, in partnership with Public Health England (PHE); and the Microbiology Society (International Development Fund, IDF14/01). N. S. is funded through a University of Oxford/PHE Clinical Lectureship; D. W. C. is supported by the National Institute for Health Research Health Protection Research Unit (NIHR HPRU) in Healthcare Acquired Infections and Antimicrobial Resistance at Oxford University in partnership with Public Health England (PHE). Transparency declarations None to declare. Disclaimer The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR, the Department of Health or Public Health England. Supplementary data Figures S1 and S2 and Tables S1 and S2 appear as Supplementary data at JAC Online. References 1 Queenan AM , Bush K. Carbapenemases: the versatile β-lactamases. Clin Microbiol Rev 2007; 20: 440– 58. Google Scholar CrossRef Search ADS 2 Carattoli A. Plasmids and the spread of resistance. Int J Med Microbiol 2013; 303: 298– 304. Google Scholar CrossRef Search ADS 3 Deshpande LM , Rhomberg PR, Sader HS et al. Emergence of serine carbapenemases (KPC and SME) among clinical strains of Enterobacteriaceae isolated in the United States Medical Centers: report from the MYSTIC Program (1999-2005). Diagn Microbiol Infect Dis 2006; 56: 367– 72. Google Scholar CrossRef Search ADS 4 Cai JC , Zhou HW, Zhang R et al. Emergence of Serratia marcescens, Klebsiella pneumoniae, and Escherichia coli isolates possessing the plasmid-mediated carbapenem-hydrolyzing β-lactamase KPC-2 in intensive care units of a Chinese hospital. Antimicrob Agents Chemother 2008; 52: 2014– 8. Google Scholar CrossRef Search ADS 5 Tsakris A , Voulgari E, Poulou A et al. In vivo acquisition of a plasmid-mediated bla(KPC-2) gene among clonal isolates of Serratia marcescens. J Clin Microbiol 2010; 48: 2546– 9. Google Scholar CrossRef Search ADS 6 Schulz-Stubner S , Zimmer P, Leonards P et al. Colonization-outbreak of two clonally different strains of Serratia marcescens in a neonatal intensive care unit. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz 2015; 58: 190– 6. Google Scholar CrossRef Search ADS 7 Aucken HM , Pitt TL. Antibiotic resistance and putative virulence factors of Serratia marcescens with respect to O and K serotypes. J Med Microbiol 1998; 47: 1105– 13. Google Scholar CrossRef Search ADS 8 Voelz A , Mueller A, Gillen J et al. Outbreaks of Serratia marcescens in neonatal and pediatric intensive care units: clinical aspects, risk factors and management. Int J Hyg Environ Health 2010; 213: 79– 87. Google Scholar CrossRef Search ADS 9 Yoon HJ , Choi JY, Park YS et al. Outbreaks of Serratia marcescens bacteriuria in a neurosurgical intensive care unit of a tertiary care teaching hospital: a clinical, epidemiologic, and laboratory perspective. Am J Infect Control 2005; 33: 595– 601. Google Scholar CrossRef Search ADS 10 Gastmeier P. Serratia marcescens: an outbreak experience. Front Microbiol 2014; 5: 81. Google Scholar CrossRef Search ADS 11 Mahlen SD , Morrow SS, Abdalhamid B et al. Analyses of ampC gene expression in Serratia marcescens reveal new regulatory properties. J Antimicrob Chemother 2003; 51: 791– 802. Google Scholar CrossRef Search ADS 12 Milatovic D , Schmitz FJ, Verhoef J et al. Activities of the glycylcycline tigecycline (GAR-936) against 1, 924 recent European clinical bacterial isolates. Antimicrob Agents Chemother 2003; 47: 400– 4. Google Scholar CrossRef Search ADS 13 Stock I , Grueger T, Wiedemann B. Natural antibiotic susceptibility of strains of Serratia marcescens and the S. liquefaciens complex: S. liquefaciens sensu stricto, S. proteamaculans and S. grimesii. Int J Antimicrob Agents 2003; 22: 35– 47. Google Scholar CrossRef Search ADS 14 Johnson AP , Woodford N. Global spread of antibiotic resistance: the example of New Delhi metallo-β-lactamase (NDM)-mediated carbapenem resistance. J Med Microbiol 2013; 62: 499– 513. Google Scholar CrossRef Search ADS 15 Kumarasamy KK , Toleman MA, Walsh TR et al. Emergence of a new antibiotic resistance mechanism in India, Pakistan, and the UK: a molecular, biological, and epidemiological study. Lancet Infect Dis 2010; 10: 597– 602. Google Scholar CrossRef Search ADS 16 Gruber TM , Goettig S, Mark L et al. Pathogenicity of pan-drug-resistant Serratia marcescens harbouring bla(NDM-1). J Antimicrob Chemother 2015; 70: 1026– 30. 17 Jain A , Hopkins KL, Turton J et al. NDM carbapenemases in the United Kingdom: an analysis of the first 250 cases. J Antimicrob Chemother 2014; 69: 1777– 84. Google Scholar CrossRef Search ADS 18 Khalifa HO , Soliman AM, Ahmed AM et al. High carbapenem resistance in clinical gram-negative pathogens isolated in Egypt. Microb Drug Resist 2017; 23: 838– 44. Google Scholar CrossRef Search ADS 19 Timofte D , Panzaru CV, Maciuca IE et al. Active surveillance scheme in three Romanian hospitals reveals a high prevalence and variety of carbapenamase-producing Gram-negative bacteria: a pilot study, December 2014 to May 2015. Euro Surveill 2016; 21: 5– 12. Google Scholar CrossRef Search ADS 20 Centers for Disease Prevention and Control. Laboratory Protocol for Detection of Carbapenem-Resistant or Carbapenemase-Producing, Klebsiella spp. and E. coli from Rectal Swabs. https://www.cdc.gov/hai/pdfs/labsettings/klebsiella_or_ecoli.pdf. 21 Dallenne C , Da Costa A, Decre D et al. Development of a set of multiplex PCR assays for the detection of genes encoding important β-lactamases in Enterobacteriaceae. J Antimicrob Chemother 2010; 65: 490– 5. Google Scholar CrossRef Search ADS 22 Nordmann P , Poirel L, Carrer A et al. How to detect NDM-1 producers. J Clin Microbiol 2011; 49: 718– 21. Google Scholar CrossRef Search ADS 23 Wood DE , Salzberg SL. Kraken: ultrafast metagenomic sequence classification using exact alignments. Genome Biol 2014; 15: R46. Google Scholar CrossRef Search ADS 24 Nguyen LT , Schmidt HA, von Haeseler A et al. IQ-TREE: a fast and effective stochastic algorithm for estimating maximum-likelihood phylogenies. Mol Biol Evol 2015; 32: 268– 74. Google Scholar CrossRef Search ADS 25 Lunter G , Goodson M. Stampy: a statistical algorithm for sensitive and fast mapping of Illumina sequence reads. Genome Res 2011; 21: 936– 9. Google Scholar CrossRef Search ADS 26 Stoesser N , Sheppard AE, Pankhurst L et al. Evolutionary history of the global emergence of the Escherichia coli epidemic clone ST131. MBio 2016; 7: e02162. Google Scholar CrossRef Search ADS 27 Tavare S. Some probabilistic and statistical problems in the analysis of DNA sequences. Lectures on Mathematics in the Life Sciences 1986; 17: 57– 86. 28 Didelot X , Wilson DJ. ClonalFrameML: efficient inference of recombination in whole bacterial genomes. PLoS Comput Biol 2015; 11: e1004041. Google Scholar CrossRef Search ADS 29 Letunic I , Bork P. Interactive tree of life (iTOL) v3: an online tool for the display and annotation of phylogenetic and other trees. Nucleic Acids Res 2016; 44: W242– 5. Google Scholar CrossRef Search ADS 30 Bankevich A , Nurk S, Antipov D et al. SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing. J Comput Biol 2012; 19: 455– 77. Google Scholar CrossRef Search ADS 31 Carattoli A , Zankari E, Garcia-Fernandez A et al. In silico detection and typing of plasmids using PlasmidFinder and plasmid multilocus sequence typing. Antimicrob Agents Chemother 2014; 58: 3895– 903. Google Scholar CrossRef Search ADS 32 Siguier P , Perochon J, Lestrade L et al. ISfinder: the reference centre for bacterial insertion sequences. Nucleic Acids Res 2006; 34: D32– 6. Google Scholar CrossRef Search ADS 33 Stoesser N , Batty EM, Eyre DW et al. Predicting antimicrobial susceptibilities for Escherichia coli and Klebsiella pneumoniae isolates using whole genomic sequence data. J Antimicrob Chemother 2013; 68: 2234– 44. Google Scholar CrossRef Search ADS 34 Orlek A , Phan H, Sheppard AE et al. A curated dataset of complete Enterobacteriaceae plasmids compiled from the NCBI nucleotide database. Data Brief 2017; 12: 423– 6. Google Scholar CrossRef Search ADS 35 Loman NJ , Quinlan AR. Poretools: a toolkit for analyzing nanopore sequence data. Bioinformatics 2014; 30: 3399– 401. Google Scholar CrossRef Search ADS 36 Berlin K , Koren S, Chin CS et al. Assembling large genomes with single-molecule sequencing and locality-sensitive hashing. Nat Biotechnol 2015; 33: 623– 30. Google Scholar CrossRef Search ADS 37 Walker BJ , Abeel T, Shea T et al. Pilon: an integrated tool for comprehensive microbial variant detection and genome assembly improvement. PLoS One 2014; 9: e112963. Google Scholar CrossRef Search ADS 38 Kearse M , Moir R, Wilson A et al. Geneious Basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics 2012; 28: 1647– 9. Google Scholar CrossRef Search ADS 39 Huang TW , Wang JT, Lauderdale TL et al. Complete sequences of two plasmids in a blaNDM-1-positive Klebsiella oxytoca isolate from Taiwan. Antimicrob Agents Chemother 2013; 57: 4072– 6. Google Scholar CrossRef Search ADS 40 Pal C , Bengtsson-Palme J, Kristiansson E et al. Co-occurrence of resistance genes to antibiotics, biocides and metals reveals novel insights into their co-selection potential. BMC Genomics 2015; 16: 964. Google Scholar CrossRef Search ADS 41 Willis LM , Whitfield C. KpsC and KpsS are retaining 3-deoxy-D-manno-oct-2-ulosonic acid (Kdo) transferases involved in synthesis of bacterial capsules. Proc Natl Acad Sci USA 2013; 110: 20753– 8. Google Scholar CrossRef Search ADS 42 Liang W , Xie Y, Xiong W et al. Anti-restriction protein, KlcAHS, promotes dissemination of carbapenem resistance. Front Cell Infect Microbiol 2017; 7: 150. Google Scholar CrossRef Search ADS 43 Szekely E , Damjanova I, Janvari L et al. First description of bla(NDM-1), bla(OXA-48), bla(OXA-181) producing Enterobacteriaceae strains in Romania. Int J Med Microbiol 2013; 303: 697– 700. Google Scholar CrossRef Search ADS 44 Deshpande LM , Flonta M, Jones RN et al. Detection of NDM-1-producing Enterobacteriaceae in Romania: report of the SENTRY Antimicrobial Surveillance Program. J Med Microbiol 2014; 63: 483– 4. Google Scholar CrossRef Search ADS 45 Dortet L , Flonta M, Boudehen YM et al. Dissemination of carbapenemase-producing Enterobacteriaceae and Pseudomonas aeruginosa in Romania. Antimicrob Agents Chemother 2015; 59: 7100– 3. Google Scholar CrossRef Search ADS 46 European Centre for Disease Prevention and Control. Antimicrobial Resistance Surveillance in Europe 2012. Annual Report of the European Antimicrobial Resistance Surveillance Network (EARS-Net), Stockholm, 2013. 47 Grundmann H , Glasner C, Albiger B et al. Occurrence of carbapenemase-producing Klebsiella pneumoniae and Escherichia coli in the European survey of carbapenemase-producing Enterobacteriaceae (EuSCAPE): a prospective, multinational study. Lancet Infect Dis 2017; 17: 153– 63. Google Scholar CrossRef Search ADS 48 Czobor I , Novais A, Rodrigues C et al. Efficient transmission of IncFIIY and IncL plasmids and Klebsiella pneumoniae ST101 clone producing OXA-48, NDM-1 or OXA-181 in Bucharest hospitals. Int J Antimicrob Agents 2016; 48: 223– 4. Google Scholar CrossRef Search ADS 49 Lixandru BE , Cotar AI, Straut M et al. Carbapenemase-producing Klebsiella pneumoniae in Romania: a six-month survey. PLoS One 2015; 10: e0143214. Google Scholar CrossRef Search ADS 50 Tijet N , Richardson D, MacMullin G et al. Characterization of multiple NDM-1-producing Enterobacteriaceae isolates from the same patient. Antimicrob Agents Chemother 2015; 59: 3648– 51. Google Scholar CrossRef Search ADS 51 Wu W , Feng Y, Carattoli A et al. Characterization of an Enterobacter cloacae strain producing both KPC and NDM carbapenemases by whole-genome sequencing. Antimicrob Agents Chemother 2015; 59: 6625– 8. Google Scholar CrossRef Search ADS 52 Wailan AM , Paterson DL, Kennedy K et al. Genomic characteristics of NDM-producing Enterobacteriaceae isolates in Australia and their blaNDM genetic contexts. Antimicrob Agents Chemother 2016; 60: 136– 41. Google Scholar CrossRef Search ADS 53 Mantilla-Calderon D , Jumat MR, Wang T et al. Isolation and characterization of NDM-positive Escherichia coli from municipal wastewater in Jeddah, Saudi Arabia. Antimicrob Agents Chemother 2016; 60: 5223– 31. Google Scholar CrossRef Search ADS © The Author 2017. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
Journal of Antimicrobial Chemotherapy – Oxford University Press
Published: Mar 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.
All for just $49/month
It’s easy to organize your research with our built-in tools.
All the latest content is available, no embargo periods.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud